After seven generations of selection, a line of mice selected for post-weaning (21–42 days) weight gain on full feeding (SF) showed significant increases of 49% in weight gain, 31% in efficiency and 14% in food intake, when compared with its control on full feeding between 21 and 42 days. After day 42, SF mice continued to eat more food and were 28% heavier than control mice at 91 days. Because SF mice were heavier than control mice at almost all ages, they were fatter on an age basis. There was, however, no change in the rate of deposition of fat, protein and ash relative to body weight. On restricted feeding between 21 and 42 days, SF mice showed a non-significant increase in weight gain, and hence in efficiency, of 12%. They deposited more fat than control mice during the feeding period but there was no significant difference when comparisons were made on a weight basis.

A contemporary line of mice selected for post-weaning (21–42 days) weight gain on restricted feeding (SR) had significant increases of 12% in weight gain, 17% in efficiency but no significant change in food intake, when compared with its control on full feeding between 21 and 42 days. SR mice were the same weight as control mice at all ages except day 21, when they were significantly lighter due to direct genetic effects rather than maternal effects. SR mice had a lower (P<0·10) rate of fat deposition per unit body weight and became less fat relative to their control as body weight increased. The rate of deposition of other components was not altered by selection. On restricted feeding, SR had a significant increase in weight gain, and hence in efficiency, of 37%. Changes in body composition were similar to those on full feeding.

It was concluded that the use of a restricted feeding regime had enabled the exploitation of heritable variation in the partitioning of energy for growth. This variation was independent of genetic variation for appetite and body weight.

Overall performance at each level of feeding was best improved by selection on that feeding level. The realized genetic correlation between post-weaning weight gain on full and restricted feeding was estimated to be 0·28 ± 0·08, indicating a very different genetic basis for the same character in the two feeding environments.

Five loci have been identified in Saccharomyces cerevisiae whose function reduces suppressor activity in strains carrying ochre super-suppressor mutations. Recessive mutations which allow an increased level of suppression occur at these loci. In such mutants, termed allosuppressors, the serine-inserting suppressor SUPQ5 suppresses ochre mutations in a [psi−] background and Class I tyrosine-inserting suppressors are lethal or have a reduced viability. Mutations at two allosuppressor loci, sal3 and sal4, have a lethal interaction with one another and with the extrachromosomal determinant [psi+]. This interaction is expressed in the absence of any suppressor mutation. All the mutant alleles of one allosuppressor locus sal3 are cold sensitive. One allosuppressor mutation, sal4.2, is temperature-sensitive for growth, as well as for other aspects of its phenotypic expression; namely the expression of SUPQ5 and the lethal interactions with Class I super-suppressors, with [psi+] and with sal3. At low temperature (24 °C), sal3-sal4.2 double mutants weakly suppress the ochre mutation ade2.1, but do not suppress his5.2 or lys1.1. It is argued that the site of function of the products of these loci is ribosomal and that they are involved in chain termination at UAA codons. It is inferred that the [psi+] factor or its product affects protein synthesis by interaction with the ribosome.

A computer simulation of the genetic analysis of the asexual life cycle of Phycomyces' heterokaryons has been carried out. We have studied experimentally the relationship between the nuclear proportion in heterokaryotic mycelia containing prototrophic and auxotrophic nuclei and their growth rates. We discuss possible evolutionary implications of heterokaryosis in coenocytic fungi and the genetic applications of the quantitative complementation technique.

A study of crosses between CBA/FaCam and C57B1/6Fa mice revealed an effect of the origin of the Y-chromosome on testis weight and aggressive behaviour, but failed to reveal any effect on sexual behaviour and androgen metabolism. There is therefore no evidence that androgens mediate the Y-linked variation in aggressive behaviour and testis weight. On behavioural grounds, it is difficult to compare measures of sexual and aggressive behaviour, but it appears from these results that there are major genetic components on the Y-chromosome controlling the development of sexual and aggressive behaviour which are distinct.

Phenethyl alcohol is shown to induce cytoplasmic respiratory deficiency (petite mutation) in various strains of Saccharomyces cerevisiae. Apart from petite induction, phenethyl alcohol also affects cell growth both in respiratile and non-respirable media. This sensitivity to growth inhibition appears to be complex, and does not correlate with the degree of induction of respiratory deficiency.

The first part of the paper provides strong supportive evidence for the previous findings (Cummins & Day, 1973; Day & Cummins, 1973) that the two alleles of the mating-type locus of the basidiomycete Ustilago violacea have different periods of inducibility during a cell cycle, and that the cell cycle characteristics of each allele are maintained in freshly isolated diploids. This difference in temporal properties of the alleles appears to be the basis of the dominance of allele a2 as it is inducible during a phase of the cell cycle when allele a1 is non-inducible. During G1 both alleles appear to be inducible and apparently ‘neutralize’ each other so that the cell cannot mate.

The second part of the paper provides evidence for a unique genetic control mechanism. The evidence suggests that the period of cell cycle inducibility of a locus governing a morphogenetic pathway may be regulated by a separate control gene the cc locus, with two known alleles ccstr(a stringent or restricted period of inducibility) and ccrel (a relaxed or non-restricted period of inducibility). This hypothesis stems from analysis of a diploid that was a1· ccstr/a2· ccrel and showed dominance of allele a2 during the S and G2 phases when freshly isolated, but which became incapable of mating after a period of subculturing. Analysis of haploids derived from this diploid strain showed that both mating-type alleles were functional but that it was now homozygous for ccstr, i.e. of genotype a1· ccstr/a2·ccstr· Thus the temporal and functional aspects of the mating type alleles are determined by different loci. It is postulated that cell cycle control loci may be widespread and serve to regulate the action of genes concerned with morphogenesis in relation to other cell cycle events.

A method was developed to estimate effects of quantitative trait loci (QTL) by maximum likelihood using information from changes of gene frequency at marker loci under selection, assuming an additive model of complete linkage between markers and QTL. The method was applied to data from 16 molecular and coat colour marker loci in mouse lines derived from the F2 of two inbred strains which were divergently selected on 6-week weight for 21 generations. In 4 regions of the genome, marker allele frequencies were more extreme than could be explained by sampling, implying selection at nearby QTL. An effect of about 0·5 standard deviations was located on chromosome 11, and accounted for nearly 10% of the genetic variance in the base population. QTL with effects as small as 0·2 phenotypic standard deviations could be detected. For typing of a given number of individuals, the power of detection of QTL is very high compared to, for example, analysis of an F2 population. The joint effects of linkage and selection were investigated by Monte Carlo simulation. Marker gene frequencies change little as a consequence of selection at a QTL unless the marker and QTL are less than about 20 cM apart.

By performing matings using various Hfr strains having different origins of transfer, it has been observed that male chromosomal markers located very near the origin of transfer in conjugation in E. coli appear in recombinants with a lower frequency than expected from an extrapolation of the transfer gradient. The effect is greatest for markers closest to the origin, and becomes negligible for markers transferred more than 4–5 min. after the origin. The results suggest that in the zygote a random cross-over between male and female genomes is necessary somewhere between the origin and any male marker in order for that marker to appear in a viable recombinant. It was also deduced that the entry time for the origin itself occurs after a 3–5 min. delay after mixing male and female cells together.

The Lac plasmids in two Klebsiella strains, V9A and RE1544, give characteristic Lac-permease defective mutations which produce a lactose-negative (ML−) phenotype in the host strain, in spite of the presence of a lac operon on the chromosome. These ML− clones can revert to the wild-type (ML+) phenotype, and the mechanism of this reversion is examined. In V9A carrying its own Lac plasmid (FKlac), it is shown that reversion of ML− to ML+ is usually the result of mutation to constitutivity of one of two galactoside permeases which are not induced by lactose but accumulate lactose when otherwise induced or made constitutive. However, in one out of the 51 ML+ revertants tested the mechanism of reversion appeared to be a change back to wild type of the Lac plasmid's own permease gene. In V9A carrying the Lac plasmid (pRE6) from RE1544, successive changes of phenotype were obtained of ML+ to ML− to ML+ to ML− to ML+; these were all found to be the result of changes in the plasmid permease gene, and could be simply explained if an IS sequence could insert and excise from this gene.

1. Forty-nine auxotrophic mutants were isolated after irradiation with ultraviolet light. The biochemical requirements were identified and the response to possible precursors tested in some of them.

2. Diploid colonies were synthesized from compatible, auxotrophic haploid strains on an artificial medium by an adaptation of the balanced heterokaryon technique. Selection of diploid cells was especially convenient as the dikaryon cannot grow on such media, and is therefore only a transitory stage under these cultural conditions. Diploid cells were different in shape and size from haploids, and gave rise to colonies which could be distinguished by eye from haploids. The way in which diploid colonies arose from fusions between haploid cells was followed by microscopic observation. When known numbers of fused haploid cells were plated on minimal medium, diploid colonies occurred at a frequency of 3 × 10−4.

3. Diploid cells, heterozygous for mating-type alleles, were incompatible with either haploid mating-type (i.e. neutral).

4. Diploid cells could infect the host plant as a pure culture (i.e. they are solopathogenic).

5. Meiotic and mitotic segregation of the large-celled strains was used to confirm diploidy. Spontaneous mitotic segregation was very rare. Mitotic crossing-over was induced at a high frequency by irradiation of diploid cells with ultraviolet light. A convenient technique for induced haploidization was devised using p-fluoro-phenylalanine. Preliminary evidence using this technique indicates a haploid chromosome number of at least seven.

6. It is considered that U. violacea has many advantages as an organism for genetical research, especially for analysis of the parasexual cycle.

1. Heterozygous (K/k) killer paramecia containing kappa particles from stock 51, syngen 4 of Paramecium aurelia were passed through autogamy, and the kk offspring tested for killing ability and for presence of stainable kappa particles at various fission stages.

2. After eight fissions, a small proportion of the paramecia were non-killers, and a few lacked stainable particles. After the 12th fission nearly all cells lacked kappa.

3. Mild starvation of the kk kappa-bearing paramecia resulted in complete loss of kappa and conversion to non-killers, even when applied to animals only one fission after the loss of gene K at autogamy.

4. The results obtained here show some discrepancy with data obtained earlier, both with kappa, and with mu particles in stock 540, syngen 1. The reasons for this discrepancy are at present not clear.

The behaviour of the X- and Y-borne Sts locus has been studied in male and female mice. There was considerable heterogeneity in STS activity between inbred mouse strains, with a four fold difference in activity between the highest (101/H) and lowest (Ju/Ct) activity strains, which can be interpreted in terms of allelic differences. In all inbred strains male STS levels were higher than those of female STS levels and in the majority of strains tested male STS levels were nearly twice as high as female levels. Reciprocal crosses between C3H/HeH and the STS-deficient substrain, C3H/An, demonstrated that activities of the X- and Y-borne genes in males are essentially the same and this suggested that the lower STS level in females derives from X-inactivation of the locus. The possibility that hormonal differences could instead be responsible for the lower activity in females was ruled out by the findings that (a) castration of males did not reduce their STS levels and (b) sex-reversed males, X / X Sxr, had STS levels typical of females. Final proof that the mouse Sts locus can be subject to the X-inactivation process was provided by the observation that XX females had STS levels that were only slightly (20%) higher than those of XO females. The difference may indicate incomplete inactivation of the locus. Linkage data verifying the location of Sts on the distal end of the X chromosome are provided.

In total, the results of this study show that the murine Sts locus can be subject to the X-inactivation process and this, together with the existence of functional loci of near-equal activities on the X and Y chromosomes, results in an imbalance of STS levels between the sexes. X-inactivation does not therefore serve as a dosage compensation mechanism for the Sts locus in the mouse. All of these findings were made in C3H/HeH mice or in animals carrying C3H/HeH functional Sts alleles, and it is pointed out that the diverse results previously obtained by other investigators may be attributable to their use of different strains and crosses between strains but could also be complicated by technical factors.

1. Individual genes of a multigenic system of resistance to actidione in yeast were analysed in a series of crosses involving resistant strains and sensitive strains.

2. Resistance conferred varied in the range 0·5 p.p.m. to 20 p.p.m.

3. Positive interaction was seen between resistance genes in recombinant strains.

4. Recessive modifier genes were found which do not themselves confer resistance but which modify the activity of specific resistance genes in a positive way, that is towards a greater degree of resistance. Both a two-fold and a five-fold increase in resistance is seen depending on the gene combination.

5. In selection experiments, strains resistant to relatively high concentrations of actidione (1000 p.p.m.) were obtained but only by selection on media containing successively higher concentrations of the drug, that is, by the step wise build up of multiply mutant forms. Mutations of interacting resistance genes and specific modifier genes were obtained by selection methods including pretreatment of cells with ultra-violet light.

1. Several lines of Drosophila melanogaster have been selected for increase of sternopleural chaeta-number.

2. Three lines of the same origin, dp 1, dp 2 and dp 6, showed remarkably similar patterns of response involving an accelerated response from 24 to 28 chaetae and a plateau at about 30 chaetae. A line formed by crossing two of these was not responsive to selection, suggesting that the two parent lines were genetically similar.

3. A fourth line, vg 4, related to these others, showed a similar accelerated response but continued to respond beyond 30 chaetae, reaching a plateau at about 37 chaetae.

4. A further line, vg 6, set up by crossing two ‘improved’ lines, one with 30, the other with 22 chaetae, reached 46 chaetae and suggests there may sometimes be merit in selecting from the hybrids of ‘improved’ strains.

5. It is argued that the event making possible the accelerated response is probably a recombinational event.

6. The line vg 4 clearly demonstrates that some of the genes that may be exploited in a selection experiment have to be exploited in a particular order because of gene interactions.

Maximum likelihood methods for the mapping of quantitative trait loci (QTL) have been investigated in an F2 population using simulated data. The use of adjacent (flanking) marker pairs gave improved power for the detection of QTL over the use of single markers when markers were widely spaced and the QTL effect large. The use of flanking marker loci also always gave moreaccurate and less biassed estimates for the effect and position of the QTL and made the method less sensitive to violations of assumptions, for example non-normality of the distribution. Testing the hypothesis of a linked QTL against that of no QTL is not biassed by the presence of unlinked QTL. This test is more robust and easier to obtain than the comparison of a linked with an unlinked QTL. Fixing the recombination fraction between the markers at an incorrect value in the analyses with flanking markers does not generally bias the test for QTL or estimates of their effect. The presence of multiple linked QTL bias both tests and estimates of effect with interval mapping, leading to inflated values when QTL are in association in the lines crossed and deflated values when they are in dispersion.